This R01 application proposes to test a theory-guided intervention to increase the likelihood that physicians will apply expert-recommended "best practices" in assessing and treating aggressive children in psychiatric inpatient and day treatment settings. Given concerns about potential side effects of the powerful agents (especially "atypical" antipsychotic medications such as risperidone, olanzapine, etc.) used to treat these children and the lack of previous research addressing this problem, studies of intervention methods to modify physicians' practices are an urgent public health necessity. Unfortunately, few physician-targeted behavioral change interventions have been developed that draw upon basic behavioral and cognitive science (BBCS) theory. As a result, most successful physician behavior change efforts across all of medicine have relied principally on intensive procedures (e.g., academic detailing) or practical combinations of multiple strategies to achieve a meaningful effect. What has not emerged from efforts to date is an adequate understanding of which behavior change strategy works for which physicians, how, and why. A theory-guided approach to this problem may not only yield increased physician application of recommended treatment practices, but could also lead to the development of more cost-effective physician behavior change interventions. To address these challenges this application proposes to test a novel intervention based on the Theory of Reasoned Action and the Theory of Planned Behavior and recent theory extensions known as "mental contrasting" and forming "implementation intentions" (MC/II), in order to assist physicians in applying a newly developed clinical practice guideline (CPG). 120 physicians will be randomly assigned to an information-only control (IOC) group, a 12-month-long academic detailing intervention group (ADI), or a similarly intense (in length of time and number of physician contacts) novel intervention (MC/II) group, with all participants followed for an additional 12 months after the active intervention has concluded. The study addresses following specific aims: Aim 1: To test the impact and benefits of a BBCS-guided intervention on physicians' behaviors related to their assessment, monitoring, and use of antipsychotic medications in severely aggressive youths, comparing this intervention to two control groups, a practice guidelines information-only control group (IOC) and an academic detailing intervention group (ADI). Aim 2: To determine the durability and maintenance of intervention effects up to one year after both active interventions have ceased, comparing the ADI and BBCS interventions with each other and with the IOC group. Aim 3: To explore the mediators and moderators of the interventions' relative effectiveness, in order to better understand which type of intervention works for which physicians under which circumstances. NY Office of Mental Health has agreed to delay the statewide rollout of the CPG until summer, 2003. This application constitutes a time-critical opportunity for carrying out the proposed research.